A major component of the pathophysiology in schizophrenia involves an imbalance between excitatory glutamate and inhibitory GABA in the cortex, and consequent problems in coherent local circuit function and at the NMDA receptor related to neuronal plasticity and learning. Coherent local circuit activity is crucial for real tie pattern analysis in sensory cortex and for rule learning involving frontal, temporal, and parietal cortices. Mismatch negativity (MMN) is a brainwave that is sensitive to Glu-GABA imbalance, and is abolished by drugs that block NMDA. MMN to simple stimulus changes (a soft tone among repetitive loud tones, or a high note among repetitive low notes) arises in primary and initial secondary auditory cortices, and is reduced in chronic schizophrenia. Yet the simple MMN is healthy at first psychotic break, though it declines during the early disease course in conjunction with primary auditory cortex gray matter loss, and is also normal in relatives. Because of this simple MMN cannot be used for family/genetic studies or for pre- or pro-dromal identification. Recently it has been discovered that MMN is elicited by more complex stimulus patterns and learned rules. These higher-order MMNs have not been studied in schizophrenia. Because of their complexity, abstraction of these rules requires involvement of sophisticated brain circuits spanning secondary cortices across frontal and temporal lobes. Complex MMN, intimately tied to NMDA-modulated memory formation, is likely to be more sensitive to and provide a more precise index of Glu-GABA imbalance in schizophrenia, even at or before first break, and in relatives. The main aim of this project is to detect the presence of complex memory MMN deficits in tasks that require second-order analysis of stimulus patterns and learned rules. As a first step, this project aims to identify complex MMN deficits in well characterized schizophrenia participants and first break patients before moving to at risk persons and relatives. Nine experiments will examine increasingly more complex second-order memory. If 5 tones played with a short delay between them are always followed by a long delay, gestalt proximity makes one form a unit of 5 tones. Violating that rule by adding or subtracting from the group of 5 elicits a complex MMN. Violations of second order rules of increasing pitch, duration, or loudness trends generate a MMN. Experience with language is used to aid perception of phonemes. (Consider how one needs to hear a foreign language for some time before perceiving the gap between words.) We predict that schizophrenics will not use learned rules for phoneme categories and will actually be more sensitive to acoustic differences in syllables than will controls. Demonstrating defects in MMN to second order memory is important for understanding the basic pathophysiology in schizophrenia, will provide novel and more sensitive indices of Glu-GABA imbalance in the disorder, and increase the utility of MMN as a tool for early identification and treatment of at risk persons before they have a full blown psychotic break, and for family studies and discovery of the basic genetic risk factors for the disease.